Am doing a comparison between classification models on forecasting the return direction. Can I use the accuracy metric from the confusion matrix and k fold cross-validation, or
